Enhanced product family delivers face recognition and computer vision from Edge to Cloud, including new toolsets for mobile, advanced liveness, large-scale search, and beyond

SAN FRANCISCO — Paravision, the U.S.-based leader in mission-critical face recognition and computer vision, today announced an array of new and updated products that make face recognition solutions faster, more accurate and secure, and easier to develop and deploy. This comprehensive product update includes broad and deep toolkits for implementing face recognition and computer vision at the Edge, on mobile devices, on-premises, and in the Cloud.

Paravision is powering partners to scale up with its latest search engine technology, Scaled Vector Search, as well as a major update to its Streaming Video Container. It is unlocking the Edge with updates to its Face SDK for Ambarella CVflow, the introduction of Ambarella-based Liveness and Anti-Spoofing, and the launch of an Ambarella Reference Design Kit with Teknique. It is powering mobile with its Face SDK for Android, and it is going beyond face recognition with person-level computer vision. Together, the new toolsets make integrating face recognition into applications across access control, payments, identity, and video security far more scalable and more cost-effective.

Scaling Up

  • Scaled Vector Search (SVS): Rapid Search for Large-Scale Identity Databases 
  • Streaming Container 5: Processes More Video Frames while Increasing Accuracy

The ability to process very large databases is critical for the successful implementation of face recognition in use cases such as identity verification, payments, and government ID programs. Paravision’s new Scaled Vector Search (SVS) capability enables accurate, rapid, and stable face matching on very large databases of identities. With extremely low latencies, the SVS engine is able to handle hundreds of transactions per second and databases of tens of millions of identities. SVS is Docker-based, cloud-ready, and can be scaled to meet any operational need.

Since processing video at scale is critical, Paravision has also introduced Streaming Container 5. Built to handle over 250 frames per second from any number of video streams, it includes advanced face tracking, ensuring that identities aren’t confused, even in busy environments. 

Unlocking the Edge

  • Face SDK for Ambarella CVflow: Higher Than Ever Accuracy for Face Recognition
  • Liveness and Anti-spoofing SDK for Ambarella CVflow: Ensures a Human Presence
  • Edge AI Reference Designs: Fast, Low-risk Development

Paravision has enhanced its Face SDK for smart cameras powered by Ambarella CVflow chipsets. This allows its face recognition to achieve up to 40 frames per second full pipeline performance on Ambarella SoCs, delivering a breakthrough combination of speed, power, accuracy, and cost-effectiveness.

Paravision is also launching a Liveness and Anti-spoofing SDK to add new safeguards for face recognition using Ambarella smart cameras. This toolkit uses Ambarella’s supported visible light, near-infrared and depth-sensing capabilities to determine whether the camera sees a live subject or is being tricked by recorded footage or a dummy image, unlocking the power of CVflow for applications that demand real-time liveness and anti-spoofing capability.

To facilitate faster, lower risk development of Edge AI solutions, Paravision has partnered with Teknique, the global leader in the development of Ambarella-based designs. The two companies have co-created a series of hardware and software reference designs that can enable rapid, high-confidence development of face recognition and computer vision capabilities at the Edge. Teknique System on Modules (SOMs), camera boards and optical designs, combined with Paravision’s SDKs make the development of AI-powered face recognition systems at the Edge more cost-effective and accessible than ever.


  • Paravision Face SDK for Android: Powerful Face Recognition on Mobile Devices

From identity verification to visitor management to national and civil ID programs, mobile devices have become a platform of choice for biometric identity. With the release of Paravision Face SDK for Android, partners can develop mission-critical mobile applications with Paravision face recognition. This full-fledged SDK includes face detection, landmarks, quality assessment, template creation, and 1-to-1 or 1-to-many matching. To facilitate rapid, derisked development, Paravision’s Android SDK comes with detailed and thoughtful reference applications which include UI/UX recommendations and tools. 

Beyond Face Recognition

  • Person Detection: Detects the Presence and Position of Individuals Unlocking New Use Cases

Complementing its face recognition technology, Paravision is announcing the availability of its first person-level computer vision SDK. Person Detection is a function for determining the presence and position of individuals in a field of view. Person Detection can be used as the basis for a number of computer vision capabilities such as occupancy analysis, the ability to spot tailgating, as well as custom intention or subject attributes. For example, with Person Detection users could determine whether employees are allowed access to a specific area, are wearing a mask or hard hat, or appear to be in distress. It can also enable useful business insights such as metrics about queue times, customer throughput or to detect traveler bottlenecks.

From Cloud to Edge, Paravision’s goal is to help our partners develop and deploy transformative solutions around face recognition and computer vision. With these sweeping updates to our product family, and with what has become possible in terms of accuracy, speed, usability and portability, we see a remarkable opportunity to unite disparate applications with a coherent sense of identity that bridges physical spaces and cyberspace.
Joey Pritikin, Paravision Chief Product Officer

Over the past year, Paravision has demonstrated its leadership in a wide range of NIST’s gold standard Face Recognition Vendor Test (FRVT) evaluations. The company has consistently achieved #1 U.S. ranking and top-10 global performance in all critical leaderboards including 1:1 verification, 1:N identification, performance for paperless travel, and performance with face masks.

Paravision’s new capabilities were unveiled at its Converge 2021 virtual conference, accessible online here. The first event of its kind, Converge addresses technology partners in AI and computer vision, chipset developers, camera and optical equipment manufacturers, as well as leaders in sectors from financial services, government, automotive, payments, physical security, and building access. All Paravision’s face recognition products are developed and sold under the guidance of its AI Principles which stipulate that its AI systems must be ‘ethically trained and conscientiously sold.’ 


Watch full video announcement below: